API

Every jigsawstack.api.* operation is listed below with parameter shapes and return types from the plugin Zod schemas.

New to Corsair? See API access, authentication, and error handling.

Ai

imageGeneration

ai.imageGeneration Generate an image from a text prompt Risk: write

await corsair.jigsawstack.api.ai.imageGeneration({});

Input

Name	Type	Required	Description
`prompt`	`string`	Yes	—
`aspect_ratio`	`string`	No	—
`width`	`number`	No	—
`height`	`number`	No	—
`steps`	`number`	No	—
`output_format`	`png \| svg`	No	—
`return_type`	`url \| binary \| base64`	No	—
`url`	`string`	No	—
`file_store_key`	`string`	No	—
`advance_config`	`object`	No	—

advance_config full type

{
  negative_prompt?: string,
  guidance?: number,
  seed?: number
}

Output: object

Output full type

{
  success: boolean,
  content_type: string,
  base64: string
} | {
  success: boolean,
  _usage?: {
    input_tokens?: number,
    output_tokens?: number,
    inference_time_tokens?: number,
    total_tokens?: number
  },
  log_id?: string,
  url?: string,
  base64?: string
}

prediction

ai.prediction Forecast a time series from dated values Risk: read

await corsair.jigsawstack.api.ai.prediction({});

Input

Name	Type	Required	Description
`dataset`	`object[]`	Yes	—
`steps`	`number`	No	—

dataset full type

{
  value: number | string,
  date: string
}[]

Output

Name	Type	Required	Description
`success`	`boolean`	Yes	—
`_usage`	`object`	No	—
`log_id`	`string`	No	—
`prediction`	`object[]`	No	—

_usage full type

{
  input_tokens?: number,
  output_tokens?: number,
  inference_time_tokens?: number,
  total_tokens?: number
}

prediction full type

{
  value?: number | string,
  date?: string
}[]

sentiment

ai.sentiment Analyze sentiment and emotion in text Risk: read

await corsair.jigsawstack.api.ai.sentiment({});

Input

Name	Type	Required	Description
`text`	`string`	Yes	—

Output

Name	Type	Required	Description
`success`	`boolean`	Yes	—
`_usage`	`object`	No	—
`log_id`	`string`	No	—
`sentiment`	`object`	No	—

_usage full type

{
  input_tokens?: number,
  output_tokens?: number,
  inference_time_tokens?: number,
  total_tokens?: number
}

sentiment full type

{
  emotion?: string,
  sentiment?: string,
  score?: number,
  sentences?: {
    text?: string,
    emotion?: string,
    sentiment?: string,
    score?: number
  }[]
}

summary

ai.summary Summarize text or a PDF as a paragraph or bullet points Risk: read

await corsair.jigsawstack.api.ai.summary({});

Input

Name	Type	Required	Description
`text`	`string`	No	—
`url`	`string`	No	—
`file_store_key`	`string`	No	—
`type`	`text \| points`	No	—
`max_points`	`number`	No	—
`max_characters`	`number`	No	—

Output

Name	Type	Required	Description
`success`	`boolean`	Yes	—
`_usage`	`object`	No	—
`log_id`	`string`	No	—
`summary`	`string \| string[]`	No	—

_usage full type

{
  input_tokens?: number,
  output_tokens?: number,
  inference_time_tokens?: number,
  total_tokens?: number
}

translate

ai.translate Translate text into a target language Risk: read

await corsair.jigsawstack.api.ai.translate({});

Input

Name	Type	Required	Description
`text`	`string \| string[]`	Yes	—
`target_language`	`string`	Yes	—
`current_language`	`string`	No	—

Output

Name	Type	Required	Description
`success`	`boolean`	Yes	—
`_usage`	`object`	No	—
`log_id`	`string`	No	—
`translated_text`	`string \| string[]`	No	—

_usage full type

{
  input_tokens?: number,
  output_tokens?: number,
  inference_time_tokens?: number,
  total_tokens?: number
}

Audio

createVoiceClone

audio.createVoiceClone Clone a voice from an audio sample for later TTS Risk: write

await corsair.jigsawstack.api.audio.createVoiceClone({});

Input

Name	Type	Required	Description
`name`	`string`	Yes	—
`url`	`string`	No	—
`file_store_key`	`string`	No	—

Output

Name	Type	Required	Description
`success`	`boolean`	Yes	—
`_usage`	`object`	No	—
`log_id`	`string`	No	—
`voice_id`	`string`	No	—

_usage full type

{
  input_tokens?: number,
  output_tokens?: number,
  inference_time_tokens?: number,
  total_tokens?: number
}

speechToText

audio.speechToText Transcribe audio or video to text Risk: read

await corsair.jigsawstack.api.audio.speechToText({});

Input

Name	Type	Required	Description
`url`	`string`	No	—
`file_store_key`	`string`	No	—
`language`	`string`	No	—
`translate`	`boolean`	No	—
`by_speaker`	`boolean`	No	—
`webhook_url`	`string`	No	—
`batch_size`	`number`	No	—
`chunk_duration`	`number`	No	—
`word_timestamps`	`boolean`	No	—

Output

Name	Type	Required	Description
`success`	`boolean`	Yes	—
`_usage`	`object`	No	—
`log_id`	`string`	No	—
`text`	`string`	No	—
`chunks`	`any[]`	No	—
`status`	`string`	No	—
`id`	`string`	No	—
`language_detected`	`any`	No	—

_usage full type

{
  input_tokens?: number,
  output_tokens?: number,
  inference_time_tokens?: number,
  total_tokens?: number
}

textToSpeech

audio.textToSpeech Convert text to speech audio Risk: write

await corsair.jigsawstack.api.audio.textToSpeech({});

Input

Name	Type	Required	Description
`text`	`string`	Yes	—
`accent`	`string`	No	—
`voice_clone_id`	`string`	No	—
`speaker_clone_url`	`string`	No	—
`speaker_clone_file_store_key`	`string`	No	—

Output

Name	Type	Required	Description
`success`	`boolean`	Yes	—
`content_type`	`string`	Yes	—
`base64`	`string`	Yes	—

Classification

classify

classification.classify Classify text or images with custom labels Risk: read

await corsair.jigsawstack.api.classification.classify({});

Input

Name	Type	Required	Description
`dataset`	`object[]`	Yes	—
`labels`	`object[]`	Yes	—
`multiple_labels`	`boolean`	No	—

dataset full type

{
  type: text | image,
  value: string
}[]

labels full type

{
  key?: string,
  type: text | image,
  value: string
}[]

Output

Name	Type	Required	Description
`success`	`boolean`	Yes	—
`_usage`	`object`	No	—
`log_id`	`string`	No	—
`predictions`	`(string \| string[])[]`	No	—

_usage full type

{
  input_tokens?: number,
  output_tokens?: number,
  inference_time_tokens?: number,
  total_tokens?: number
}

Embedding

createV2

embedding.createV2 Create v2 embeddings from text, image, audio, or PDF Risk: read

await corsair.jigsawstack.api.embedding.createV2({});

Input

Name	Type	Required	Description
`text`	`string`	No	—
`url`	`string`	No	—
`file_store_key`	`string`	No	—
`type`	`text \| text-other \| image \| audio \| pdf`	Yes	—
`token_overflow_mode`	`truncate \| error`	No	—
`dimensions`	`number`	No	—
`instruction`	`string`	No	—
`query`	`boolean`	No	—
`speaker_fingerprint`	`boolean`	No	—

Output

Name	Type	Required	Description
`success`	`boolean`	Yes	—
`_usage`	`object`	No	—
`log_id`	`string`	No	—
`embeddings`	`number[][]`	No	—
`chunks`	`any`	No	—
`speaker_embeddings`	`number[][]`	No	—

_usage full type

{
  input_tokens?: number,
  output_tokens?: number,
  inference_time_tokens?: number,
  total_tokens?: number
}

Prompt Engine

create

promptEngine.create Create a reusable Prompt Engine template Risk: write

await corsair.jigsawstack.api.promptEngine.create({});

Input

Name	Type	Required	Description
`prompt`	`string`	Yes	—
`name`	`string`	No	—
`return_prompt`	`any`	No	—
`inputs`	`object[]`	No	—
`use_internet`	`boolean`	No	—
`optimize_prompt`	`boolean`	No	—
`prompt_guard`	`defamation \| specialized_advice \| privacy \| intellectual_property \| indiscriminate_weapons \| hate \| sexual_content \| elections \| code_interpreter_abuse[]`	No	—

inputs full type

{
  key: string,
  optional?: boolean,
  initial_value?: string
}[]

Output

Name	Type	Required	Description
`success`	`boolean`	Yes	—
`_usage`	`object`	No	—
`log_id`	`string`	No	—
`prompt_engine_id`	`string`	No	—

_usage full type

{
  input_tokens?: number,
  output_tokens?: number,
  inference_time_tokens?: number,
  total_tokens?: number
}

list

promptEngine.list List Prompt Engine templates Risk: read

await corsair.jigsawstack.api.promptEngine.list({});

Input

Name	Type	Required	Description
`page`	`number`	No	—
`limit`	`number`	No	—

Output

Name	Type	Required	Description
`success`	`boolean`	Yes	—
`_usage`	`object`	No	—
`log_id`	`string`	No	—
`prompt_engines`	`any[]`	No	—
`page`	`number`	No	—
`limit`	`number`	No	—
`has_more`	`boolean`	No	—

_usage full type

{
  input_tokens?: number,
  output_tokens?: number,
  inference_time_tokens?: number,
  total_tokens?: number
}

run

promptEngine.run Run a stored Prompt Engine by id Risk: write

await corsair.jigsawstack.api.promptEngine.run({});

Input

Name	Type	Required	Description
`id`	`string`	Yes	—
`input_values`	`object`	No	—
`stream`	`boolean`	No	—

input_values full type

{
}

Output

Name	Type	Required	Description
`success`	`boolean`	Yes	—
`_usage`	`object`	No	—
`log_id`	`string`	No	—
`result`	`any`	No	—
`message`	`string`	No	—

_usage full type

{
  input_tokens?: number,
  output_tokens?: number,
  inference_time_tokens?: number,
  total_tokens?: number
}

Validate

nsfw

validate.nsfw Detect NSFW content in an image Risk: read

await corsair.jigsawstack.api.validate.nsfw({});

Input

Name	Type	Required	Description
`url`	`string`	No	—
`file_store_key`	`string`	No	—

Output

Name	Type	Required	Description
`success`	`boolean`	Yes	—
`_usage`	`object`	No	—
`log_id`	`string`	No	—
`nsfw`	`boolean`	No	—
`nudity`	`boolean`	No	—
`gore`	`boolean`	No	—
`nsfw_score`	`number`	No	—
`nudity_score`	`number`	No	—
`gore_score`	`number`	No	—

_usage full type

{
  input_tokens?: number,
  output_tokens?: number,
  inference_time_tokens?: number,
  total_tokens?: number
}

profanity

validate.profanity Check text for profanity and return a cleaned copy Risk: read

await corsair.jigsawstack.api.validate.profanity({});

Input

Name	Type	Required	Description
`text`	`string`	Yes	—
`censor_replacement`	`string`	No	—

Output

Name	Type	Required	Description
`success`	`boolean`	Yes	—
`_usage`	`object`	No	—
`log_id`	`string`	No	—
`message`	`string`	No	—
`clean_text`	`string`	No	—
`profanities_found`	`boolean`	No	—
`profanities`	`object[]`	No	—

_usage full type

{
  input_tokens?: number,
  output_tokens?: number,
  inference_time_tokens?: number,
  total_tokens?: number
}

profanities full type

{
  profanity?: string | null,
  startIndex?: number,
  endIndex?: number
}[]

spamCheck

validate.spamCheck Score text for spam likelihood Risk: read

await corsair.jigsawstack.api.validate.spamCheck({});

Input

Name	Type	Required	Description
`text`	`string \| string[]`	Yes	—

Output

Name	Type	Required	Description
`success`	`boolean`	Yes	—
`_usage`	`object`	No	—
`log_id`	`string`	No	—
`check`	`object[]`	No	—

_usage full type

{
  input_tokens?: number,
  output_tokens?: number,
  inference_time_tokens?: number,
  total_tokens?: number
}

check full type

{
  is_spam?: boolean,
  score?: number
} | {
  is_spam?: boolean,
  score?: number
}[]

spellCheck

validate.spellCheck Detect and auto-correct spelling mistakes Risk: read

await corsair.jigsawstack.api.validate.spellCheck({});

Input

Name	Type	Required	Description
`text`	`string`	Yes	—
`language_code`	`string`	No	—

Output

Name	Type	Required	Description
`success`	`boolean`	Yes	—
`_usage`	`object`	No	—
`log_id`	`string`	No	—
`misspellings_found`	`boolean`	No	—
`auto_correct_text`	`string`	No	—
`misspellings`	`object[]`	No	—

_usage full type

{
  input_tokens?: number,
  output_tokens?: number,
  inference_time_tokens?: number,
  total_tokens?: number
}

misspellings full type

{
  word?: string | null,
  startIndex?: number,
  endIndex?: number,
  expected?: string[],
  auto_corrected?: boolean
}[]

Vision

detectObjects

vision.detectObjects Detect objects and GUI elements in an image Risk: read

await corsair.jigsawstack.api.vision.detectObjects({});

Input

Name	Type	Required	Description
`url`	`string`	No	—
`file_store_key`	`string`	No	—
`prompts`	`string[]`	No	—
`enhance_prompts`	`boolean`	No	—
`features`	`object \| gui[]`	No	—
`annotated_image`	`boolean`	No	—
`return_type`	`url \| base64`	No	—
`return_masks`	`boolean`	No	—
`return_tags`	`boolean`	No	—

Output

Name	Type	Required	Description
`success`	`boolean`	Yes	—
`_usage`	`object`	No	—
`log_id`	`string`	No	—
`annotated_image`	`string`	No	—
`gui_elements`	`any[]`	No	—
`objects`	`any[]`	No	—
`tags`	`string[]`	No	—

_usage full type

{
  input_tokens?: number,
  output_tokens?: number,
  inference_time_tokens?: number,
  total_tokens?: number
}

vocr

vision.vocr Extract text and fields from an image or PDF Risk: read

await corsair.jigsawstack.api.vision.vocr({});

Input

Name	Type	Required	Description
`url`	`string`	No	—
`file_store_key`	`string`	No	—
`prompt`	`object`	No	—
`page_range`	`number[]`	No	—
`fine_grained`	`boolean`	No	—
`return_bounds`	`boolean`	No	—

prompt full type

string | string[] | {
}

Output

Name	Type	Required	Description
`success`	`boolean`	Yes	—
`_usage`	`object`	No	—
`log_id`	`string`	No	—
`context`	`any`	No	—
`width`	`number`	No	—
`height`	`number`	No	—
`tags`	`string[]`	No	—
`has_text`	`boolean`	No	—
`sections`	`any[]`	No	—
`total_pages`	`number`	No	—

_usage full type

{
  input_tokens?: number,
  output_tokens?: number,
  inference_time_tokens?: number,
  total_tokens?: number
}

Web

htmlToAny

web.htmlToAny Convert HTML or a URL to PNG, JPEG, WEBP, or PDF Risk: write

await corsair.jigsawstack.api.web.htmlToAny({});

Input

Name	Type	Required	Description
`html`	`string`	No	—
`url`	`string`	No	—
`type`	`pdf \| png \| jpeg \| webp`	No	—
`full_page`	`boolean`	No	—
`width`	`number`	No	—
`height`	`number`	No	—
`return_type`	`url \| binary \| base64`	No	—
`quality`	`number`	No	—
`is_mobile`	`boolean`	No	—
`dark_mode`	`boolean`	No	—
`size_preset`	`string`	No	—

Output: object

Output full type

{
  success: boolean,
  content_type: string,
  base64: string
} | {
  success: boolean,
  _usage?: {
    input_tokens?: number,
    output_tokens?: number,
    inference_time_tokens?: number,
    total_tokens?: number
  },
  log_id?: string,
  url?: string,
  base64?: string
}

scrape

web.scrape Scrape a page into structured data with AI prompts Risk: read

await corsair.jigsawstack.api.web.scrape({});

Input

Name	Type	Required	Description
`url`	`string`	No	—
`html`	`string`	No	—
`element_prompts`	`object`	No	—
`selectors`	`string[]`	No	—
`root_element_selector`	`string`	No	—
`scroll`	`boolean`	No	—
`page_position`	`number`	No	—
`features`	`meta \| link[]`	No	—

element_prompts full type

string[] | {
}

Output

Name	Type	Required	Description
`success`	`boolean`	Yes	—
`_usage`	`object`	No	—
`log_id`	`string`	No	—
`data`	`any[]`	No	—
`page_position`	`number`	No	—
`page_position_length`	`number`	No	—
`meta`	`any`	No	—
`link`	`any[]`	No	—
`selectors`	`object`	No	—
`context`	`any`	No	—

_usage full type

{
  input_tokens?: number,
  output_tokens?: number,
  inference_time_tokens?: number,
  total_tokens?: number
}

selectors full type

{
}

search

web.search Search the web with optional AI overview Risk: read

await corsair.jigsawstack.api.web.search({});

Input

Name	Type	Required	Description
`query`	`string`	Yes	—
`spell_check`	`boolean`	No	—
`max_results`	`number`	No	—
`safe_search`	`strict \| moderate \| off`	No	—
`ai_overview`	`boolean`	No	—
`auto_scrape`	`boolean`	No	—
`country_code`	`string`	No	—
`byo_urls`	`string[]`	No	—

Output

Name	Type	Required	Description
`success`	`boolean`	Yes	—
`_usage`	`object`	No	—
`log_id`	`string`	No	—
`query`	`string`	No	—
`ai_overview`	`string`	No	—
`spell_fixed`	`boolean`	No	—
`is_safe`	`boolean`	No	—
`results`	`any[]`	No	—
`image_urls`	`string[]`	No	—
`links`	`string[]`	No	—
`geo_results`	`any[]`	No	—

_usage full type

{
  input_tokens?: number,
  output_tokens?: number,
  inference_time_tokens?: number,
  total_tokens?: number
}

searchSuggestions

web.searchSuggestions Get search autocomplete suggestions for a query Risk: read

await corsair.jigsawstack.api.web.searchSuggestions({});

Input

Name	Type	Required	Description
`query`	`string`	Yes	—

Output

Name	Type	Required	Description
`success`	`boolean`	Yes	—
`_usage`	`object`	No	—
`log_id`	`string`	No	—
`suggestions`	`string[]`	No	—

_usage full type

{
  input_tokens?: number,
  output_tokens?: number,
  inference_time_tokens?: number,
  total_tokens?: number
}

Overview

Plugins

Ai

imageGeneration

prediction

sentiment

summary

translate

Audio

createVoiceClone

speechToText

textToSpeech

Classification

classify

Embedding

createV2

Prompt Engine

create

list

run

Validate

nsfw

profanity

spamCheck

spellCheck

Vision

detectObjects

vocr

Web

htmlToAny

scrape

search

searchSuggestions

​Ai

​imageGeneration

​prediction

​sentiment

​summary

​translate

​Audio

​createVoiceClone

​speechToText

​textToSpeech

​Classification

​classify

​Embedding

​createV2

​Prompt Engine

​create

​list

​run

​Validate

​nsfw

​profanity

​spamCheck

​spellCheck

​Vision

​detectObjects

​vocr

​Web

​htmlToAny

​scrape

​search

​searchSuggestions

Ai

imageGeneration

prediction

sentiment

summary

translate

Audio

createVoiceClone

speechToText

textToSpeech

Classification

classify

Embedding

createV2

Prompt Engine

create

list

run

Validate

nsfw

profanity

spamCheck

spellCheck

Vision

detectObjects

vocr

Web

htmlToAny

scrape

search

searchSuggestions